1
\$\begingroup\$

I've built a multiple-feedback low-pass filter with a TL084IN op amp. I've simulated the filter for the chosen R and C values using LTSpice and it looks like I was expecting: LTSpice simulated Bode Plot.

I've also assembled the filter in real life and checked the Bode Plot using NIMyDAQ. This time I get a lot of noise once the frequency goes past 1k, which is the part that I don't understand: weird NI MyDAQ Bode Plot.

Any ideas why MyDAQ's Bode looks like this ?

\$\endgroup\$
  • 1
    \$\begingroup\$ 1) that's not "multiple feedback", it's still one feedback loop even though is has multiple feedback paths. 2) the ripple you see in the phase plot probably changes for each measurement right ? So it's a random behavior. I would not call it noise though. What happens at these higher frequencies ? What does the output signal do, you can see that in the gain plot. The signal gets very small it probably gets so small that your measurement device cannot determine the phase correctly anymore. So it just shows rubbish instead. \$\endgroup\$ – Bimpelrekkie May 18 '16 at 19:10
  • \$\begingroup\$ Do you have adequate bypass caps on the power supply pins?. You may also be exceeding the limits of your ADC. What type of card did you get for LabView? \$\endgroup\$ – Sparky256 May 18 '16 at 21:14
  • \$\begingroup\$ @FakeMoustache It is a filter topology known as a Multiple Feedback (MFB) Filter. \$\endgroup\$ – user207421 May 19 '16 at 0:30
2
\$\begingroup\$

You don't have a whole lot of signal there at the higher attenuations so I think those are just artifacts of the measurement.

If the behavior deep into the stop band of your MFB filter is really of keen interest, try to increase the signal levels to as high as your ADC can handle without clipping. You have only 16 bits so the the quantization noise will limit you around -95dB. There is probably significantly more noise than that in the ADC. If that doesn't work well enough for your requirements, get a better ADC or even try a sound card.

\$\endgroup\$
1
\$\begingroup\$

You are misinterpreting your graphs. A wildly-swinging phase response is not oscillation. I suggest you try providing a nominal zero volts DC in, and do a transient response analysis of the output, and I suspect you'll see no oscillation. If there is any, it will be at microvolt levels, and will be an artifact of the simulator.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.